AI Process Management Visualization for Non-Deterministic Workflows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional BPM tools fail to effectively visualize and manage AI-based process flows due to the non-deterministic nature of artificial intelligence implementations, leading to complexity and inefficiency in software development, with users becoming frustrated and requiring extensive IT resources without achieving significant improvements.
Innovation Solution
A system and method that integrates project management tools with AI-based business process management, allowing non-technical users to model and manage AI-enabled processes using a human-resource-based paradigm, enabling seamless transition to technical resources for development and monitoring, with a graphical user interface that displays task and rule information, and utilizes machine learning to predict AI system actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional BPM tools are used to visualize and manage AI-based process flows, then the tools provide basic process management functionality, but they fail to effectively handle the non-deterministic nature of AI implementations, leading to increased complexity and frustration
Solution Approach 1:
The patent introduces an intermediary layer between the deterministic BPM tools and non-deterministic AI processes. This intermediary consists of probabilistic task representations and predictive modeling components that translate AI's uncertain outputs into manageable process states, allowing conventional BPM tools to effectively manage AI-driven workflows without direct exposure to their non-deterministic nature
Solution Approach 2:
The patent transforms the parameters used to represent task states from deterministic values to probabilistic distributions. By changing how task completion states are parameterized (from binary complete/not-complete to probability distributions), the system can accommodate AI's non-deterministic outputs while maintaining process manageability and reducing overall system complexity
2Ease of operation
If non-technical users are enabled to model and manage AI processes, then ease of operation improves, but the complexity of predicting and monitoring AI system actions increases
Solution Approach 1:
The patent implements feedback mechanisms where machine learning models continuously predict AI system actions and update task probability states based on actual outcomes. This feedback loop allows non-technical users to interact with simplified interfaces while the system automatically handles the complexity of predicting and monitoring AI actions through iterative learning and state updates
Solution Approach 2:
The patent replaces manual prediction and monitoring mechanisms with automated machine learning models. Instead of requiring users to manually track and predict AI system actions, the system uses ML algorithms to automatically detect, measure, and predict AI behaviors, substituting complex mechanical tracking with intelligent automation that simplifies user interaction
3Reliability
If IT resources are extensively deployed to manage AI processes, then process management capability improves, but development costs and resource requirements increase significantly
Solution Approach 1:
The patent enables the system to manage itself through automated machine learning models that continuously monitor, predict, and adjust task states without requiring extensive manual IT intervention. The self-service capability allows the BPM system to autonomously handle the complexity of AI process management, maintaining high reliability while minimizing the quantity of IT resources needed
Solution Approach 2:
The patent changes the parameter representation from detailed deterministic states to compact probabilistic distributions, reducing the information processing burden. This parameter transformation allows the system to maintain accurate process management capability with fewer resources by working with compressed, probability-based state representations rather than exhaustive deterministic details
Data Source
AI summary
Aspects of the subject disclosure may include, for example, providing a graphical user interface on a display device for interaction with operations personnel associated with a process to be performed; receiving, from the operations personnel, information defining one or more tasks to perform the process; receiving, from technical personnel associated with respective task of the one or more tasks, respective rulesets associated with the respective tasks, each respective ruleset defining procedures to complete a respective task, one or more of the respective rulesets implementing a machine learning algorithm to complete the respective task; displaying information about the one or more tasks on the graphical user interface during performance of the process; estimating, with an artificial intelligence process, a current status of the process, producing a current status estimate, wherein the estimating comprises estimating a status of the machine learning algorithm; and displaying process status information on the graphical user interface wherein the process status information is based on the current status estimate. Other embodiments are disclosed.